Dark Kitchen and Virtual Brand Unit Economics: Viability Modeling and Per-Channel Cost Structure

Chapter 1 — When is a dark kitchen actually viable in 2026?

A dark kitchen is viable in 2026 only if its prime cost stays healthy after deducting the aggregator's commission and it runs 2-3 virtual brands over a single kitchen line.

The mistake I see again and again working with operators: modeling margin on gross menu price instead of the effective ticket. Delivery is no longer marginal: it has become one of the largest revenue pools in global food service. That structural shift turns the kitchen-without-a-dining-room into a low-CapEx expansion model. But the global ghost kitchen market will hit USD 204 billion by 2030 according to GlobeNewswire (2026), and most units fail from poorly modeled unit economics, not lack of demand. Demand exists; margin has to be built with discipline. That distinction is the whole game.

Chapter 2 — Net price is not menu price

A dark kitchen's net price is not its menu price: after the aggregator's commission, which depends on the deal, and promotional discounts, the effective ticket drops by a sizable fraction. This is the #1 error I see in new operators: they project revenue on gross price and discover too late that a dish priced well on the menu leaves far less in the till. Delivery volume is enormous, but that volume isn't yours: the platform intermediates it. DoorDash moved an enormous volume of orders for local merchants in 2024, and Just Eat Takeaway processed EUR 26.3 billion in GTV (Just Eat Takeaway.com 2024). Billing heavily on a poorly modeled gross ticket means subsidizing the aggregator with your own working capital.

Chapter 3 — Why is prime cost measured after commission?

A dark kitchen's prime cost is measured after the aggregator's commission, never before. A food cost that looks perfectly healthy becomes unsustainable when a large slice of revenue goes to the platform before it touches your till.

The math is unforgiving: if the aggregator takes 25% and your gross food cost is 30%, on net revenue that real food cost climbs to ~40%, and adding production labor the prime cost crosses 60% easily. At Masterestaurant we always model from net revenue per order, not from the menu. The market pushes that way: the ghost kitchen market is growing at an 11.65% CAGR between 2022 and 2032, according to Statista/Toast (via OysterLink). Growing on a mis-counted prime cost only accelerates cash burn. Contribution per order after commission is the only figure that decides whether you open or close.

Chapter 4 — Multi-brand is a margin lever, not vanity

Multi-brand is a margin lever, not a vanity exercise: running 2 to 4 virtual brands over the same kitchen dilutes CapEx and fixed OpEx per order, but it only works if they share mise en place and suppliers. I've seen it in dozens of operations: three brands with distinct menus that actually use the same base of proteins, sauces and sides turn a one-shift kitchen into a three-stream revenue machine without adding rent or equipment. Each extra brand that shares inputs lowers the break-even of the whole kitchen. The trap is launching brands with separate purchasing: there you multiply waste instead of margin, and fixed OpEx per order rises instead of falling.

Chapter 5 — The direct channel recovers the aggregator's margin

The direct channel —your own web, WhatsApp, white-label app— is the only one that recovers the margin the aggregator takes. Without a deliberate strategy to migrate customers to the direct channel, the unit economics stays trapped in the aggregator's commission forever. Every order you move from the platform to your own channel recovers those points at once: a ticket that left you far less on the aggregator now leaves the full amount minus the payment gateway cost, which is a small fraction. The aggregator is acquisition; the direct channel is retention. The discipline of capturing the customer's data on the first order and reactivating them via WhatsApp is what separates a profitable dark kitchen from one that only fattens the platform's GTV.

Chapter 6 — Viability model: contribution per order rules

A dark kitchen's viability is decided by contribution margin per order after commission, not by gross order volume. An operator billing hard on Rappi or iFood but not controlling prime cost is subsidizing the aggregator with working capital. The equation I use at Masterestaurant is simple and hard: net revenue per order minus real food cost minus production labor must leave positive contribution before touching rent and utilities, which go to break-even, not to the dish. The sector confirms the scale: DoorDash closed 2024 with ~USD 80.2 billion in marketplace GOV (DoorDash 2024). But that GTV belongs to the platform. Your business lives or dies on unit contribution: if it's negative, every extra order pushes you toward bankruptcy faster. Model the worst input scenario first; if it holds there, it holds anywhere.

Chapter 7 — The differences that decide viability

Net price is not menu price: after the aggregator's commission, which varies by contract, and promotional discounts, the effective ticket drops by a sizable fraction. Modeling on gross price is mistake #1 among new operators. Dark kitchen prime cost is measured AFTER commission, not before. A food cost that looks healthy on paper becomes unsustainable when a large slice of revenue goes to the platform before it touches your cash. Multi-brand is a margin lever, not vanity: 2-4 virtual brands on one kitchen dilute fixed CapEx and OpEx per order, but only if they share mise en place and suppliers. The direct channel (own web, WhatsApp, white-label app) is the only one that recovers the aggregator's margin. Without a customer-migration strategy to the direct channel, unit economics stalls.